1. Field of the Invention
This invention relates to a switching power source means having a DC voltage source with at least two switch elements connected thereto, so as to produce a DC output, or an AC output at a desired frequency, by turning on and off the switch elements in an alternate manner. In particular, the invention relates to prevention of surge currents and switching losses due to parasitic capacitors in the switching power source means.
2. Related Art Statement
A switching power source means of the above-mentioned type is generally small and highly efficient, and has been widely used, for instance, as a power source for a computer data processing system. In the case of a power source means with a large output capacity, or an AC power source means producing sinusoidal AC output from a DC power source, the power source means is often made controllable by connecting paired switch elements thereto so as to regulate the power or frequency of its output simply by alternately changing over the operating state of the switching elements.
FIG. 6 shows a circuit diagram of a typical example of a conventional controllable power source means. The waveforms at different points in the circuit thereof are shown in FIGS. 7A and 7B. Paired switch elements, i.e., a first switch element 1 and a second switch element 2, are joined in series at a point 7 and serially connected across a DC voltage source 5. A low-pass filter, which is formed of a choke coil 3 and a capacitor 4, is connected to the DC voltage source 5 through the first switch element 1. The input end of the low-pass filter is connected in parallel with the second switch element 2. A load 6 is connected in parallel with the output end of the low-pass filter. If switch elements 1 and 2 are alternately turned on and alternately turned off, while changing the ratio between ON-time and OFF-time of the switch element 2 in a sinusoidal manner, a series of pulse width modulated (PWM) voltage pulses are generated at the point 7 as shown by the curve in FIG. 7A. After elimination of switching frequency component and harmonics from the series of the voltage pulses by the low-pass filter made of the choke coil 3 and the capacitor 4, a sinusoidal AC output voltage can be obtained as shown by the curve in FIG. 7B.
The above functional description of the switching power source means of FIG. 6 is based on the assumptions that the switch elements 1 and 2 perform an exactly rectangular ON-OFF switching operation as ideal switch elements can do, and the control signals for driving the ON-OFF operation are also exactly rectangular. Performance of actual switch elements to be used, however, deviates from the above-mentioned exact rectangular switching operation due to intrinsic characteristics of individual switch elements, which causes certain difficulties during switching operation.
The difficulties during the switching operation will be explained by referring to a typical switch element of a metal oxide semiconductor field effect transistor (MOS-FET) of FIG. 8. MOS-FET's are quite frequently used as the switch elements 1 and 2 of FIG. 6. The circuit of FIG. 8 is essentially the same as that of FIG. 6 except that MOS-FETl and MOS-FET2 are used as the switch elements 1 and 2 respectively.
The MOS-FET is different from a conventional bipolar transistor in that the MOS-FET is free from delay in turn-off operation due to storage time caused by residual carriers. Thus, if control signals applied to the gates of MOS-FET1 and MOS-FET2 of FIG. 8 are exactly rectangular, simultaneous ON states of the MOS-FET1 and MOS-FET2 will never occur. However, the actual MOS-FET has a comparatively large parasitic capacitor between its drain and source on the order of about several hundred pF to several ten pF. At the time of turn-ON and turn-OFF of the paired MOS-FET1 and MOS-FET2, there is a serious problem of how to handle the electric charge stored in the parasitic capacitors so as to eliminate adverse effects of the stored charge on the turning ON and turning OFF function.
FIG. 9 shows an equivalent circuit of a MOS-FET. Due to its configuration, parasitic capacitors are inevitable; namely, C.sub.dg between the gate and the drain, C.sub.ds between the drain and the Source, and C.sub.gs between the gate and the source. The value of resistance R.sub.ds between the drain and the source varies greatly from almost zero to infinity depending on the gate-source voltage, and the zero value corresponds to the ON state and the infinity value corresponds to the OFF state. A parasitic diode D.sub.o must be considered between the drain and the source.
Phenomena relating to the turn-ON and turn-OFF of the equivalent circuit of FIG. 9 will now be explained. Electric charge stored in the drain-source parasitic capacitor C.sub.ds of each MOS-FET is discharged through the drain-source resistance R.sub.ds when it is turned ON. Hence, when the paired MOS-FET's are switched over from one to the other, a large surge current is produced. For example, a large surge current flows through the ON-state drain-source resistance R.sub.ds of the MOS-FET2 when the MOS-FET2 is turned ON from its OFF state. The surge current is due to two reasons, namely, the drain-source parasitic capacitor C.sub.ds of the MOS-FET2 is discharged through the ON-state drain-Source resistance R.sub.ds, and the drain-source parasitic capacitor C.sub.ds of MOS-FET1 is directly charged by the DC voltage source 5 through the ON-state drain-source resistance R.sub.ds of the MOS-FET2. Similarly, a large surge current flows through the ON-state drain-source resistance R.sub.ds of the MOS-FET1 when the MOS-FET2 is turned OFF and the MOS-FET1 is turned ON due to the same two reasons as described above.
As a result, the charge stored in the drain-source parasitic capacitor C.sub.ds of each MOS-FET Of the paired switch elements of FIG. 8 causes a large surge current through the ON-state drain-source resistance R.sub.ds when the two-MOS-FET's are switched over from one to another. The energy of such surge current is converted into and consumed as heat, which means not only power loss and temperature rise of the switch element but also generation of noise. Further, intensity of such phenomenon increases with the increase of the switching frequency of the paired switch elements. Thus, the occurrence of such surge current makes it very difficult to use high frequency switchover of the paired switch elements of the switching power source means. Further, if the peak value of such surge current is too high, it may cause break-down of the switch elements.
Several protective methods against the surge current accompanying the switchover of the switch elements have been proposed. FIG. 10A through FIG. 10C illustrate some of them. Gate resistors 8.sub.a1 and 8.sub.a2 of several hundred ohms are serially connected to the gates of the MOS-FET1 and MOS-FET2 as shown in FIG. 10A. Functions of such gate resistors include the reduction of the rise-rate of the gate-source voltage of each MOS-FET, the gradual change of the value of the drain-source resistance R.sub.ds of each MOS-FET at the time of switchover, and the suppression of the peak value of the surge current even if not eliminated. FIG. 10B shows the case in which two saturable magnetic cores 8.sub.b1 and 8.sub.b2 are connected in series to the MOS-FET1 and MOS-FET2, respectively. Snubber circuits 8.sub.c1 and 8.sub.c2, each comprising for instance a resistor and a capacitor, are connected in parallel to the MOS-FET1 and MOS-FET2, respectively, as shown in FIG. 10C.
The inventors found that the conventional protective methods could not ensure complete prevention of the occurrence of the above-mentioned surge current. Even if the protective methods are used, electric charge stored in the parasitic capacitors of the MOS-FET's is eventually consumed as heat in the drain-source resistance R.sub.ds. Thus, with increase of the frequency of switchover of switch elements in the switching power source means, its power consumption or heat generation increases. In short, prevention of occurrence of the surge current and reduction of switching loss at switchover of the switch elements is a very important problem to be solved in the conventional switching power source means.